1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
|
#! /usr/bin/env python
# -*- coding: utf-8 -*-
"""
Compare the speed of several methods for reading and loading a sound file.
Optionally, this file can make use of the following packages:
- audioread https://github.com/beetbox/audioread
- scipy https://scipy.org
- librosa https://github.com/bmcfee/librosa
- pydub https://github.com/jiaaro/pydub
Uncomment the function names below and send us your speed results!
"""
test_functions = [
"read_file_aubio",
"load_file_aubio",
#"load_file_scipy",
#"load_file_scipy_mmap",
#"read_file_audioread",
#"load_file_librosa",
#"read_file_pydub",
#"load_file_pydub",
]
import numpy as np
def read_file_audioread(filename):
import audioread
# taken from librosa.util.utils
def convert_buffer_to_float(buf, n_bytes = 2, dtype = np.float32):
# Invert the scale of the data
scale = 1./float(1 << ((8 * n_bytes) - 1))
# Construct the format string
fmt = '<i{:d}'.format(n_bytes)
# Rescale and format the data buffer
out = scale * np.frombuffer(buf, fmt).astype(dtype)
return out
with audioread.audio_open(filename) as f:
total_frames = 0
for buf in f:
samples = convert_buffer_to_float(buf)
samples = samples.reshape(f.channels, -1)
total_frames += samples.shape[1]
return total_frames, f.samplerate
def load_file_librosa(filename):
import librosa
y, sr = librosa.load(filename, sr = None)
#print y.mean(), y.shape
return len(y), sr
def load_file_scipy(filename):
import scipy.io.wavfile
sr, y = scipy.io.wavfile.read(filename)
y = y.astype('float32') / 32767
#print y.mean(), y.shape
return len(y), sr
def load_file_scipy_mmap(filename):
import scipy.io.wavfile
sr, y = scipy.io.wavfile.read(filename, mmap = True)
#print y.mean(), y.shape
return len(y), sr
def read_file_pydub(filename):
from pydub import AudioSegment
song = AudioSegment.from_file(filename)
song.get_array_of_samples()
return song.frame_count(), song.frame_rate
def load_file_pydub(filename):
from pydub import AudioSegment
song = AudioSegment.from_file(filename)
y = np.asarray(song.get_array_of_samples(), dtype = 'float32')
y = y.reshape(song.channels, -1) / 32767.
return song.frame_count(), song.frame_rate
def read_file_aubio(filename):
import aubio
f = aubio.source(filename, hop_size = 1024)
total_frames = 0
while True:
_, read = f()
total_frames += read
if read < f.hop_size: break
return total_frames, f.samplerate
def load_file_aubio(filename):
import aubio
f = aubio.source(filename, hop_size = 1024)
y = np.zeros(f.duration, dtype = aubio.float_type)
total_frames = 0
while True:
samples, read = f()
y[total_frames:total_frames + read] = samples[:read]
total_frames += read
if read < f.hop_size: break
assert len(y) == total_frames
#print y.mean(), y.shape
return total_frames, f.samplerate
def test_speed(function, filename):
times = []
for _ in range(10):
start = time.time()
try:
total_frames, samplerate = function(filename)
except ImportError as e:
print ("error: failed importing {:s}".format(e))
return
elapsed = time.time() - start
#print ("{:5f} ".format(elapsed)),
times.append(elapsed)
#print
times = np.array(times)
duration_min = int(total_frames/float(samplerate) // 60)
str_format = '{:25s} took {:5f} seconds avg (±{:5f}) to run on a ~ {:d} minutes long file'
print (str_format.format(function.__name__, times.mean(), times.std(), duration_min ))
if __name__ == '__main__':
import sys, time
if len(sys.argv) < 2:
print ("not enough arguments")
sys.exit(1)
filename = sys.argv[1]
for f in test_functions:
# get actual function from globals
test_function = globals()[f]
test_speed(test_function, filename)
|
